Trimmed mustache logic-less templates for python
Pull request Compare This branch is even with hyperturtle:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.gitignore
.travis.yml
LICENSE
README.md
__init__.py
setup.py
test.py

README.md

Build Status

Stache

Trimmed mustache logic-less templates

Implements everything from Mustache.5 except for lambdas in < 200 lines of code. Plus four new things. Implied closing tags {{/}}, Self referencer {{.}}, Existence check {{?exists}}{{/exists}} and data pusher {{< blah}}{{/blah}}, {{:default}}

Also, the ability to compile to javascript code!

render_js(template_string)

Compiles an inline script to javascript code

Stache().render_js_template(template_name)

Compiles all the templates and sets the entry point to the template name

render_all_js()

Compiles all the templates and returns a template object.

stachio = Stache()
stachio.add_template('template_name', templatefile.read())
write('var t = ' + stachio.render_all_js() ';')
var content = t['template_name']([{myparams:3}])
$("#container").html(content)

Why?

Because the current Pystache implementation has holes. And because I wanted to learn about python generators. As a result my codebase is considerabley smaller and easier to grok too(at least for me). It consists of two main methods, _tokenize, and _parse, both python generators. _tokenize creates tokens and _parse consumes and renders them. Also benchmarking the two with my tests, mine was slightly faster, around 2x to 3x.

Existing Stuff

{{tag}}

Renders the value of tag, html escaped, within the current scope

{{{unescape}}} & {{&unescape}}

Don't html escape the value

{{#section}}{{/section}}

Section blocks. Renders the enclosed block if

  • section is true
  • section exists

If section exists and is a(n):

  • Array: It renders the enclosed block for each element in the array, placing the current element in scope
  • Dict: It renders the enclosed block once and places the Dict as the current scope

{{^invert}}{{/invert}}

Renders the enclosed block if invert is an empty string, empty array, false, or doesn't exist. The opposite the the section block.

{{! comments - ignore me }}

Ignores the text within the tag

{{>partial}}

Looks up the partial template and renders it with the current context

New Stuff

{{/}} Implied closing tag

Whenever you use {{/}} it implies the closing of the nearest block.

{{#open}}stuff goes here{{/}}

Is the same as:

{{#open}}stuff goes here{{/open}}

{{.}} Self Referencer

This renders the current "scope". This is useful if you want to iterate over an array and wrap them.

{{#array}}<li>{{.}}</li>\n{{/array}}

with array = [1,2,3,'yay'] will produce:

<li>1</li>
<li>2</li>
<li>3</li>
<li>yay</li>

Existence Check {{?exists}}{{/}}

Forces a check of the tag name, rather than imply that it is a section block. This is useful for check if an array has members rather than iterate over the members

{{?array}}
stuff\n
{{/}}

with {array: [1, 2, 3, 4]} results in:

stuff

as opposed to

{{#array}}
stuff\n
{{/}}

which would render

stuff
stuff
stuff
stuff

{{:default}}stuff{{/}}

This is equivalent to {{default}}{{^default}}stuff{{/}}

It renders the enclosed section if default doesn't exist, empty or false

{{<thing}} Pusher {{/thing}}

It renders the inner block and adds it to the global scope.

{{<thing}}
It takes this. You can put anything in here.
{{tags}}, {{#blocks}}{{/blocks}}, etc.
{{/thing}}

and it populates the global scope with a key of thing. Watch out, it can override existing vars. A convention such as

{{<namespace.thing}}{{/namespace.thing}}

or similiar will help with collisions. This is helpful if you want to use stache templates for masterpages/inheritance. Lets say you have these templates:

master =

<div id="header">
{{header}}
</div>

<div id="footer">
{{footer}}
</div>

page =

{{<header}}
{{name}}
{{/header}}

{{<footer}}
footer
{{/footer}}

{{>master}}

Rendering the page template with {'name': 'Stachio'} will produce

<div id="header">
Stachio
</div>

<div id="footer">
footer
</div>

You can also apply the inverted block or default block to supply default blocks

master =

<div id="header">
{{header}}
{{^header}}Default Header{{/header}}
</div>

<div id="footer">
{{:footer}}Default Footer{{/footer}}
</div>

Rendering {{<footer}}Custom Footer{{/footer}}{{>master}} with {} will produce

<div id="header">
Default Header
</div>

<div id="footer">
Custom Footer
</div>

Install

pip install stache

Test

You can run python test.py or if you have nosetests:

cd stache
nosetests

Benchmark

python test.py

Usage:

>>> from Stache import Stache
>>> Stache().render("Hello {{who}}!", dict(who="World"))
Hello World

or

>>> import Stache
>>> Stache.render("Hello {{world}}!", dict(world="Stache!"))
Hello Stache!

To populate partials:

>>> from Stache import Stache
>>> stachio = Stache()
>>> stachio.add_template('main', 'a = {{a}};')
>>> stachio.add_template('main1', 'b = [ {{#b}}{{.}} {{/b}}];')
>>> stachio.add_template('main2', 'stachio')
>>> stachio.add_template('woah', '{{>main}} {{>main1}} {{>main2}}')
>>> stachio.render_template('woah',dict(a=1, b=[1,2,3,4,5]))
a = 1; b = [ 1 2 3 4 5 ] stachio

If you want to put in dynamic file loading of partials you can override Stache().templates with a dict() like object and override the __get__ and load the template in __get__ if it doesn't exist. Once you load up the template, you'll need to call self.add_template(template_name, template) to tokenize the template.

I don't think this is ideal though... Ideas for populating partials are welcome.

Efficient use with async wsgi:

For wsgi apps that support async, you can yield parts of the rendered template as they render. render_iter and render_template_iter both produce iterators that are yield'ed as it is generated.

>>> for part in Stache.render_iter("Hello {{world}}!", dict(world="Stache!")):
>>>     yield part
Hello
Stache!

Timeline:

I'm wary of lambdas, because I want the templates to be language agnostic. The main reason I liked Mustache in the first place is because of possibility of template reuse.

Some future ideas I have is rendering to javascript templates to be used on browser frontend, bypassing the need for a client side script to compile it into javascript